From “Understanding” to “Control”: A Dual-Pathway Model of How AI Interface Explainability Shapes Trust Formation in High-Risk Aviation Tasks

As artificial intelligence (AI) becomes increasingly integrated into general aviation, explainability has emerged as a critical determinant of effective human–AI collaboration and pilot trust. However, limited research has examined how explainability features influence trust formation through cognitive and psychological pathways in high-risk flight contexts. Therefore, this study proposes and validates a dual-pathway cognitive model that systematically explores how AI explainability affects user trust via two distinct psychological mechanisms. Specifically, Information Clarity influences Perceived Trust through User Engagement, while Transparent Reasoning affects Perceived Trust via Perceived Control. A structural equation modeling (SEM) approach and a controlled experiment in a simulated flight environment were employed to examine these pathways empirically. Results indicate that both explainability features significantly predict perceived trust, with the mediating roles of User Engagement and Perceived Control. Furthermore, task complexity moderates these effects, making cognitive pathways more salient under high-complexity conditions. These findings advance theoretical understanding of explainability-driven trust formation and offer actionable implications for optimizing interface transparency and user alignment mechanisms in cognitively demanding aviation scenarios.